2.15.6.1 LWR UO2 and MOX Fuels

The LWR fuel designs are described in Section 39.3.2. There are some differences in the fuel assem­bly fabrication process between PWRs and BWRs. On the other hand, there is no major difference between UO2 and MOX with respect to the fuel assembly fabrication. As an example, the flow sheet of PWR fuel assembly fabrication is shown in Figure 26.

2.15.6.1.1 Rod fabrication

The fabrication of LWR fuel rods involves the intro­duction of fuel pellets and a spring into the cladding tube, followed by welding of the end plugs and the cladding tube. For PWRs, the rods are filled with helium at a higher pressure than for BWRs. For this purpose, the top plug has a hole through which the fuel rod is pressurized, and then the hole is arc — welded. The fuel rods are inspected for surface con­tamination, dimensions, appearance, plug welds, leak tightness, and uranium enrichment. The fabrication and inspection operations are highly automated and use advanced inspection technologies, such as an X-ray image digitizing system.

2.09.4.1 FBR Application

Type 316 stainless steel was the most commonly used steel for FBR applications, and was used in the early prototype and demonstration reactors in the United States and around the world in the mid-to-late 1960s, until void swelling was discovered in 1967. As shown in Figure 15, type 316 is very prone to void swelling. Alloy D9 is an advanced austenitic alloy that was developed during the US National Cladding and Duct Development Program in the 1970s and 1980s.23 This program was designed to provide advanced materials for the liquid metal fast breeder program with a primary goal of reducing swelling at high relative to types 304 and 316 stainless steels.

D9 is a Fe-15Cr-15Ni alloy with Ti added to pro­duce TiC particles during reactor irradiation or higher temperature creep. Slight variants on this composition have been used in nuclear reactor appli­cations in the United Kingdom, France, Germany, Japan, Russia, and most recently, India. A variant of D9 has currently been used successfully as cladding and for other components in both the Phenix and SuperPhenix reactors. The D9-type austenitic stain­less steel has a clear advantage in void swelling resis­tance compared to 316 steel, but at high doses, voids form and swelling occurs (Figure 15). Several advanced austenitic stainless steels, including the creep-resistant HT-UPS steel, based on much more stable nano-dispersions of MC-precipitate micro­structures relative to the D9-type steel may have better void swelling resistance than D9

steel.10,11,14,24 However, FBR irradiation data are

needed on the HT-UPS steel to establish such benefits.

Currently, FBR technology in the United States is one of the advanced reactor options being considered by the Gen IV Nuclear Energy Systems Initiative.26 Advanced austenitic steels like D9 have higher maximum allowable design stresses for structural
components that are in the sodium-cooled reactor compared to standard 316 steel but are not exposed to the highest radiation doses found for fuel cladding and duct components. Figure 17 shows a comparison of the allowable stress benefits of the Ti-modified D9-type alloy, on the basis of higher values of UTS at lower temperatures and design rules that define maximum allowable stress as 33% of the UTS. At higher temperatures, creep-rupture strength is more limiting that tensile strength, so the maximum allow­able stress is defined as 66% of the creep-rupture stress for rupture after 100 000 h. While creep — rupture of the D9 alloy is only modestly better than that of type 316 (Figure 6), the design window defined by UTS and creep-rupture properties is larger. The HT-UPS steels are austenitic stainless steels developed from the same austenitic steel alloy composition as D9, but with a combination of addi­tions of Ti, V, and Nb rather than just Ti, and minor additions of B and P (Table 3). These composi­tional modifications to the HT-UPS steels produced unusually stable nano-dispersions of MC-carbide precipitates for much better creep-resistance than the D9 steel at 700-800 °C (Figure 6). 10,11 The creep-rupture resistance and strength of the HT-UPS steels are far superior to that of 316 and 347 steels, better than that of other advanced creep-resistant steels, such as the Nb-stabilized 347HFG, Super 304H, and NF709 austenitic stain­less steels and alloys, and comparable to that of the solid-solution strengthened Ni-based superalloy, 617, as shown in Figure 5. The creep-resistance of the HT-UPS steel at 700 °C and 170MPa is several orders of magnitude better than that of the D9 steel, as shown in Figure 6, so it should provide even larger design benefits for advanced FBR applications. Since FBR technology has recently evolved to include small, modular reactor systems as well as the more traditional larger reactor systems, advanced steels such as the HT-UPS could provide reactor designers with attractive options to improve or optimize FBR systems without dramatically increasing cost.

The CVI SiC/SiC composites develop a network of matrix cracks under load. The density of matrix cracks is enhanced by rather strong interfaces: the crack spacing may be as small as 10-20 pm whereas it is at least 10 times larger in the presence of rather

weak interfaces. Matrix cracking is an alternative mechanism of energy dissipation.

A process zone of diffused matrix microcracks is generated at the notch tip or at the tip of a preexisting main macroscopic crack. Extension of this crack results from the random failures of fiber bundles located within the process zone.35 Due to the pres­ence of a more or less large process zone associated with a jagged crack, a crack length cannot be defined and conventional concepts of fracture mechanics are not appropriate (stress intensity factor) or cannot be easily determined (strain energy release rate, У-integral). Although the validity of the stress inten­sity factor concept to measure fracture toughness is questionable, this is an interesting characteristic for comparing CVI SiC/SiC composites to other materials. Fracture toughness values on the order of 30MPaVm have been measured on single edge notch bending (SENB) test specimens.2, Strain energy

release rates ranging from 3 to 8 KJ m~2 have been determined on CVI SiC/SiC composites, respec­tively, with weak or strong interfaces.35 The cor­responding values of the У-integral ranges from 11 KJ m~ (weak interfaces) to 29 KJ m~ (strong interfaces).35 These values are quite high. The afore­mentioned stress intensity factors are maintained up to at least 1400 ° C.2

2.15.1 Introduction

Almost all the commercial nuclear power plants operating currently utilize uranium oxide fuel. These reactors, sometimes referred to as Generation II or Generation III reactors, produce ~15% of the world’s electricity supply. Production of the uranium oxide fuel required for these reactors is a mature industry and it annually requires more than 68 000 tU.1

Fuel design differs according to the reactor types, which include the advanced gas cooled reactors (AGRs), pressurized water reactors (PWRs), boiling water reactors (BWRs), PWRs developed in the for­mer Soviet Union (Vodo-Vodyanoi Energetichesky Reaktor, VVERs), and CANadian Deuterium Uranium (CANDU) reactors. There are some differences in the production processes to fit each fuel design.

Plutonium utilization within the closed fuel cycle is essential to utilize natural uranium resources effi­ciently. Plutonium recycling demonstrations have been conducted in light water reactors (LWRs) and heavy water reactors (HWRs).2 Industrial utilization of MOX in LWRs has commenced in some countries.

The use of MOX in fast neutron reactors has many attractive features. Plutonium breeding in fast breeder reactors (FBRs) leads to drastically increased energy output from uranium resources. Nuclide transmutation by fast neutrons to incinerate minor actinides (MAs) has the potential to reduce the long­term radio-toxicity of spent nuclear fuel.

2.15.2 Summary of Oxide Characteristics

2.15.2.1 Thermal and Mechanical Properties of Oxides

The starting material for oxide fuel production is oxide powder. It is fed to a powder preparation process and then to a pelletizing process to get powder compacts, which are called green pellets. The green pellets undergo a dewaxing and sintering process to get sin­tered oxide pellets. Certain characteristics of the oxide powder and the sintered pellets are very important for fuel production. A brief summary of their important characteristics is presented in this section. As a com­prehensive review of the characteristics of actinide oxide has been given in Chapter 2.02, Thermody­namic and Thermophysical Properties of the Acti­nide Oxides, most of the data presented here are those dealt with in Chapter 2.02, Thermodynamic and Thermophysical Properties ofthe Actinide Oxides.

The engineering design process starts with the utility defining its preferred refueling interval (usu­ally between 12 and 24 months depending on the grid demand profile) and the average load factor expected. This in turn determines the excess reactivity at the beginning of the cycle and the difference between this and the reactivity hold-down capability of the control rods and soluble poison (if applicable) defines the reactivity hold-down requirement for burnable poisons. The nuclear designer must then set the num­ber and location of burnable poisons to meet this requirement. The next step is to adjust the burnable poison loading in each location such that the poison material is nearly completely depleted by the end of the cycle. For IFBA poisons, this may involve choosing to use enriched 10B, while for gadolinia, it will involve choosing the optimum initial concentrations (usually in the range 4-8 wt% gadolinia). This is usually an iterative process carried out as part of the core design; it involves finding the fuel assembly loading pattern that best meets the design and safety constraints. Clearly, a compromise is required between the nuclear designer’s ideal, which might demand different num­bers of poison rods in each assembly and different initial poison concentrations, and the manufacturing constraints, which demand as simple a solution as possible.

The initial choice of burnable poison is usually dictated by the manufacturing capabilities of the fuel supplier. The IFBA design requires a major capital investment in the fuel manufacturing plant, while gadolinia rods usually have to be manufactured in a dedicated facility to avoid inadvertent contamination of the main fuel production lines (where even a few ppm of gadolinia contamination can cause fuel to go out of specification). These capital costs and the associated operational costs are passed on to the utility as part of the fuel fabrication price and appear as an indirect cost to the utility. Depending on the burnable poison type, there may also be a residual absorption penalty. For a given cycle length, this will increase the initial enrichment requirement of the fuel, which the utility sees as a direct cost.

Though large in absolute terms, relative to the overall generating costs of a commercial reactor, they are not so important. The benefit to the utility comes through being able to extend the fuel cycle length as much as possible. For a fixed refueling and mainte­nance outage time, the overall electrical output can be increased if the time between refueling outages is extended. Since fuel costs typically only represent around 20% of the overall generating cost (with capital and finance accounting for about 60%), the penalty on the fuel cost is easily outweighed by even a modest increase in output. Therein lies the main benefit and justification of burnable poisons. Absorp­tion of neutrons is generally regarded as a thing to be avoided in reactor design and operation, but burnable poisons are usually important in allowing reactor operations to be optimized. Even relatively small incre­mental improvements in reactor operations are very often sufficient to offset the negative aspects of burnable poisons, such as residual neutron absorption.

Two highly desirable properties of both neutron reflectors and moderators are efficient neutron slow­ing and low neutron absorption. The first requires effective slowing of neutrons over short distances, thus reducing the required volume of the reflector or moderator in the reactor core. Moreover, in a reactor core of a given shape and volume, this reduces the leakage of neutrons in the course of their slowing.

For reflectors in particular, the key requirements include a high reflectivity, a large macroscopic cross­section, and efficient neutron slowing. The reflectivity of a material is inversely proportional to its diffusion ratio (D/L), which is the ratio of its diffusivity (D) to its diffusion length (L). This ratio is generally considered to decrease as scattering becomes large in comparison with absorption. It is essential, moreover, to obtain high reflectivity without excessive thickness, and for this purpose, to use a material with a large macroscopic total cross-section. In a thermal reactor, the performance of the reflector is enhanced if it does not simply reflect the neutrons but rather slows and then reflects them, and for this reason, the same material is often used as both reflector and moderator.

In general, materials whose nuclides have low mass number and neutron absorption may be used as mod­erators and reflectors. The most commonly used materials are light water (H2O), heavy water (D2O), and graphite (C). In addition, hydrocarbons, zirco­nium hydride, and other such materials are often used as moderators. Heavy water is particularly effec­tive because of its very low absorption level. Graphite is second to heavy water in its low absorption level, is lower in cost, and has the added advantage of suitabil­ity for use at high temperatures. Beryllium is gener­ally used as a reflector rather than as a moderator.

In addition to the aforementioned materials, there exist other candidates as neutron reflectors. For exam­ple, Commissariat a l’Energie Atomique (CEA) is studying zirconium silicide as the reflector for next generation reactors.1 Tungsten carbide has also been used as neutron reflectors (http://en. wikipedia. org/ wiki/Tungsten_carbide). For fusion reactors, various materials such as titanium carbide and boron carbide are considered as reflectors.2

This chapter outlines the basic properties of be­ryllium and zirconium hydride that are fundamental to their utilization as neutron reflectors and modera­tors in nuclear reactors.

Typical room-temperature mechanical properties are summarized in Table 3. Measurements of micro­indentation hardness of ZrCx are prevalent in the literature. Hardness as a function of temperature is plotted in Figure 23 and as a function of the C/Zr ratio at room temperature in Figure 24. Room- temperature hardness ranges from 20 to 34 GPa (^2000-3300kgfmm~2). Hardness decreases with increasing test temperature, dropping to approxi­mately 0.5 GPa (49kgfmm~2) at 1800 K. Room — temperature hardness decreases with decreasing C/Zr ratio. Scatter in room-temperature measure­ments may be due to the variety of procedures reported (Knoop or Vickers indenter, 50-500 g load), which may not be in accordance with standard test methods.109,110 Hardness may be affected by sample microstructure, including porosity, grain morphology, and secondary phases. Residual stresses present in ion-beam deposited or pyrolytic ZrC coatings53,107 tend to inflate hardness, while free carbon reduces hardness.107,111

450

2.13.5.3 Strength and 26, respectively. Only one room-temperature

tensile strength is reported,95 and ample scatter is

Ultimate tensile strength and bend strength are

evident in room-temperature bend strength. As in

aKosolapova.95

bShaffer and Hasselman.54

cLepie.96

dGridneva et a/.97

eFedotov and Yanchur.98

fLanin et a/.99

9Neshpor eta/53

hRamqvist.8

‘Funke et a/.100

jKohlstedt.101

kSamsonov et a/.102

‘Samsonov et a/.10

mArtamonov and Bovkun.103

nAndrievskii et a/.104

oVahldiek & Mersol.105

pTkachenko et a/.106

qKumashiro et a/.123

rKumashiro et a/.124

sHe et a/.107

tWarren.90

uLanin et a/.108

covalent ceramics, ZrC fractures in an exclusively brittle manner below ^1000 K,3 by both transgranular and intergranular means. Both tensile and bend strength increase with temperature as plastic slip in­creases the resistance to brittle fracture. A maximum precedes a subsequent decrease in strength, due to decreasing yield strength with temperature, and fail­ure occurs by macroscopic plastic deformation.

The effects of porosity, grain size, specimen sur­face condition, and impurity phases remain unex­plored, with sample preparation and microstructural characteristics tending to overshadow the effects of C/Zr ratio on strength. More measurements on well — characterized samples according to standard test methods are necessary.

The fabrication process of the gadolinia-doped fuel is almost the same as that of the UO2 fuel. The gadolinia-doped fuel fabrication line must be sepa­rated from the UO2 fuel to prevent gadolinium from contaminating the UO2 fuel fabrication line.

2.15.4 MOX Production

The utilization of plutonium in reactors is essen­tial for the establishment of the nuclear fuel cycle. It is already being used in LWRs and research and development (R&D) has been continued to utilize plutonium more efficiently in FBRs. MOX fuel is often selected as FBR fuel because of its excellent burn-up potential, high melting point, and relative ease of commercial fabrication and also because LWR fuel fabricators already have extensive experi­ence with UO2 fuel fabrication. Furthermore, oxide fuel has good irradiation stability, and proven safety
response using a negative Doppler coefficient that mitigates over-power transients.42’43 These advan­tages must be weighed against the disadvantages of oxide fuel, such as lower thermal conductivity that leads to fuel structuring and enhanced swelling’44 reduced compatibility with sodium,45-47 low fissile atom density, and the presence of two moderating atoms per one metal atom. Based on a balance between the advantages and disadvantages, various fabrication processes for MOX fuels, including the conversion processes for plutonium oxide, were developed more than 40 years ago and are still applied. Major processes utilized in the conversion of plutonium oxide and MOX fuel production are summarized here. Their details have been described in the literature.2’б’27’29’42’48 Plutonium emits a-particles with energies higher than 5 MeV, and all operations from powder han­dling to end plug welding after pellets are loaded into a cladding tube are carried out in glove boxes. In order to prevent plutonium inhalation accidents during fuel fabrication, these glove boxes have an airtight structure and their interiors are continu­ously kept at negative pressure. Furthermore, as described in Section 39.2.2, gamma and neutron shielding is required for these glove boxes to reduce radiation exposure.49

Specific heat is an important parameter for the transient behavior studies where the temperature variations are linked to the variations of reactor power. Also, it is required for the calculation of thermal conductivity from thermal diffusivity. Only a very limited number of studies are available and the specific heat of irradiated fuel is not yet fully clarified. The effects of soluble fission product ele­ments added to fresh UO2 were quantified, for instance, by Verall and Lucuta,33 Matsui et al.,34 and Takahashi and Asou.35 No large burnup effect was found because the specific heat obeys the law of mixtures (Neumann-Kopp law) and because only a limited fraction of the fresh fuel heavy metal atoms change nature during irradiation.

Specific heat measurements for irradiated fuel by calorimetric techniques show an exothermic effect during the first heatup of the sample linked to the recombination of radiation damage and to fission products redistribution. The apparent specific heat is lower than for annealed samples because of the heat effect, as observed by Gomme et al?6 and Yagnik and Turnbull.37 A similar effect is observed for (U, Pu)O2 samples damaged by autoirradiation.38 This means, for instance, that during fast power increases, the temperature will increase faster than predicted using the fresh fuel-specific heat. For the intrinsic specific heat (i. e., measured on annealed
samples), no significant difference was found when compared with fresh fuel. Similar results were obtained by direct measurements of specific heat on irradiated fuels by laser flash, reported by Ronchi et al.1 for UO2 and by Sonoda et al.23 for (U, Gd)O2. Therefore, the specific heat of irradiated fuel is gen­erally assumed to be equal to that of the fresh fuel.

Nickel shows excellent corrosion resistance in caustic solutions, such as sodium hydroxide and potassium hydroxide. The corrosion rates of nickel and nickel — based alloys in sodium hydroxide are shown in Table 15.7’8’10’31

Ni shows excellent corrosion resistance to liquid sodium hydroxide at any temperature and at any concentration with water; in particular, it has an unmeasurable corrosion rate at sodium hydroxide con­centrations lower than 50%. Nevertheless, Ni corrodes in sodium hydroxide in the presence of a sulfide such as sodium sulfide. Nickel-chromium-iron alloys have excellent corrosion resistance to sodium hydrox­ide and also high resistance to ammonium hydroxide and sulfide corrosion at high temperatures. The nickel-molybdenum alloys and Alloy C, which is a nickel-chromium-molybdenum alloy, are inferior to nickel-chromium-iron alloys with respect to corrosion in ammonium hydroxide and sulfide.